207 results about "Classification types" patented technology

A method and apparatus for performing classification in a hierarchical classification system performing caching are described. In one embodiment, the method comprises walking a classification tree in the hierarchical classification system to determine whether an incoming flow matches a class in the classification tree, and performing a lookup on a cache storing a data structure of multiple classes of one classification type to compare the incoming flow with multiple classes at the same time to determine whether the incoming flow matches one of the classes.

The present invention leverages classification type detectors and / or context information to provide a systematic means to recognize and anchor annotation strokes, providing reflowable digital annotations. This allows annotations in digital documents to be archived, shared, searched, and easily manipulated. In one instance of the present invention, an annotation recognition method obtains an input of strokes that are grouped, classified, and anchored to underlying text and / or points in a document. Additional instances of the present invention utilize linguistic content, domain specific information, anchor context, and document context to facilitate in correctly recognizing an annotation.

The invention belongs to the field of image processing, and specifically discloses a chip defect recognition method based on a convolutional neural network (CNN). The method comprises the following steps of: acquiring respective types of chip images according to preset classification types, and performing size normalization to construct a data set; constructing a CNN initial model including an input layer, a convolution layer, a downsampling layer, a full connection layer and an output layer; randomly extracting images from the constructed data set to construct a training set, and inputting the images in the training set into the constructed CNN initial model for training to obtain a CNN final model; acquiring a chip image to be recognized as the input layer to be directly input into the final CNN final model, and completing the image feature extraction by successively passing the chip image through the convolution layer, the downsampling layer and the full connection layer, and enabling the output layer to output a recognition result so as to complete online chip defect recognition. The method can realize the rapid recognition of chip defects, has high recognition accuracy and high speed.

The present invention leverages classification type detectors and / or context information to provide a systematic means to recognize and anchor annotation strokes, providing reflowable digital annotations. This allows annotations in digital documents to be archived, shared, searched, and easily manipulated. In one instance of the present invention, an annotation recognition method obtains an input of strokes that are grouped, classified, and anchored to underlying text and / or points in a document. Additional instances of the present invention utilize linguistic content, domain specific information, anchor context, and document context to facilitate in correctly recognizing an annotation.

The invention discloses a deep convolutional neural network-based human face occlusion detection method. The method comprises the steps of performing block segmentation on an input image to obtain a target pre-selected region; constructing a first deep convolutional neural network, training the first deep convolutional neural network comprising a first deep convolutional network and a first multilayer perceptron connected with the first deep convolutional neural network to obtain required parameters, extracting features of the target pre-selected region, and performing classification; predicting the position of a human head through a second multilayer perceptron according to the extracted features; filtering the credibility of a classification type which is the human head and the predicted position of the human head through non-maximum suppression to remove an overlapped duplicate detection box; and obtaining a human head block in combination with original image segmentation, constructing a multi-task learning policy-based second deep convolutional neural network, and judging whether the left eye, the right eye, the nose and the mouth of the human head block are occluded or not. According to the method, the occluded human face can be accurately detected and the specific occluded part of the human face can be judged; and the method is mainly used for crime pre-warning of videos of a camera in front of an automatic teller machine.

The invention discloses a teaching quality assessment method and system based on student behavior analysis in a classroom. Through an image acquisition module, a sound acquirer, a sound pickup and an electromagnetic radiation detector, video images, speech signals of a teacher, classroom environmental noise signals and electromagnetic radiation quantity in a teaching classroom are acquired in real time and transmitted to a teaching computer for processing. The processed data is output to a server terminal for following processing, so teaching quality assessment reports are acquired. According to the invention, corresponding assessment reports are generated according to different classification types of timetables, courses, teaching contents, grades and teachers; teaching quality of the teachers and listening quality of the students are tracked in a real-time online manner; comprehensive assessment is provided for objective, quantitative and comprehensive classroom 'teaching' quality and 'listening' quality effects; and quantitative data indexes are provided for knowing and improving teaching contents of teachers afterwards and tracking of teaching quality of the school.

The invention discloses a high-resolution remote-sensing multifunctional urban land spatial information generation method. According to the method, a complex system rank theory is introduced; an urban land spatial information classification system with three rank scales, i.e. a multifunctional target landscape type, a functional area type and a land cover type, which self-adapts to urban planningmanagement and environmental renovation, is proposed; and on the basis of realizing the treatment of fine correction and alignment on remote-sensing images of a Landsat TM, a Google Earth and auxiliary maps, the Landsat TM is applied to carrying out urban landscape type classification, a three-level rank classification type-merged combination and an information mining knowledge base are constructed, the classified information is merged to form a first-level classification result of urban land, the digital functional areas are classified into a second level and the land cover is classified into a third level under the constrained control of higher-level classification information. The method has the characteristics of low cost, high accuracy of classification and strong targeted application, and thus, the requirements on target applications, such as ecological urban design, urban environmental management and the like, are better met.

The invention relates to an intelligent interaction system and method. The system includes an audio receiving module, a real-time processing module and an execution module, wherein the audio receiving module is used for receiving audio information inputted by a user, the real-time processing module is used for performing parallel online real-time processing on the audio information, and the execution module is used for executing corresponding operation according to identification results transmitted by the real-time processing module. The parallel online real-time processing includes the following steps that: classification processing and identification processing corresponding to different types are performed on the audio information; if credible classification types are obtained before the ending of audio input, identification processing on classification types except the credible classification types is terminated; identification results corresponding to the credible classification types can be obtained and are transmitted to the execution module. With the intelligent interaction system and method of the invention adopted, the user can use audio identification and voice interaction functions easily and quickly, and user experience can be enhanced.

The invention discloses an automatic regression testing method. The automatic regression testing method includes a first step, performing regression starting and running, in other words, respectively managing regression tests on different kinds of test vectors in a classified and graded manner according to specific conditions of projects, respectively selectively performing the module-level, subsystem-level or system-level regression tests for different stages of hierarchical verification and generating conventional information files and error information files; a second step, performing regression information post-processing, in other words, statistically analyzing each grade of regression test results, generating project regression home pages, generating module or regression classification branch pages and generating detailed regression result branch pages of each module. The project regression home pages contain project information, regression versions and coverage rates. The module or regression classification branch pages contain module classification type lists and pass or fail test case summaries. The detailed regression result branch pages of each module contain each test case name, simulation running time, random frequencies, case passing information, fail type statistics and simulation result conventional information file indexes. The automatic regression testing method has the advantage that the design verification process efficiency and the verification completeness can be improved by the aid of the automatic regression testing method.

The invention discloses a mammary gland molybdenum target image automatic classification method based on deep learning. The method comprises the following steps that step one, square image blocks are selected from the cancerous area and the normal area of a mammary gland molybdenum target image by using different sizes of sliding windows, and a training sample set and a test sample set corresponding to each size are constructed for different sizes of image blocks; step two, a convolutional neural network model corresponding to each size is established, and the model is trained by using the training sample set for each size; step three, the accuracy rate of the corresponding convolutional neural network model is tested by using the test sample set for each size, and the convolutional neural network model for the size corresponding to the highest accuracy rate is selected; step four, the overall connection layer characteristics are extracted by using the selected convolutional neural network model; and step five, the extracted characteristics are inputted to a linear SVM classifier for classification so that the classification types of the image blocks are obtained. The overall connection layer characteristics in the convolutional neural network model are extracted to act as the classification characteristics of the image blocks so that the classification speed and accuracy can be enhanced.

The invention provides a method and a device for saving call information of a mobile terminal and the mobile terminal. The method comprises the following steps of: receiving a call request sent by a caller by the mobile terminal or sending a call request to a caller by the mobile terminal; receiving a recording instruction of a user according to the call request; recording a call of the mobile terminal according to the recording instruction to acquire voice call information; converting the voice call information into text call information by the mobile terminal; dividing the text call information into at least one information item by the mobile terminal according to a plurality of classification types and a plurality of key words corresponding to each classification type, when the call is ended, displaying the information items in a classification manner to be selected and saved by the user, and providing an entry button of next potential operation according to content characteristics of the information items. According to the method disclosed by the embodiment of the invention, the storage space of the mobile terminal is saved, the converted text information is intelligently classified and processed to provide information to be selected and saved by the user as required and simultaneously provide the entry button of the next potential operation, so that the intelligentization of the entire process is enhanced obviously, and the user experience is promoted.

A device for aiding the flight management of an aircraft includes a navigation display, a personalized database capable of storing navigation-aid data, each datum comprising a classification type representative of a family or of one or more categories of navigation-aid data, one or more periods of validity of the said datum, information relating to at least one geographic zone of the said datum, one graphic and / or textual representation of the said datum, selection means allowing an operator to choose one (or more) type(s) of classification, called selected type(s), extraction means capable of extracting from the personalized database the navigation-aid datum or data called extracted data of which the type is equal to at least one selected type, of which at least one geographic zone is situated at a distance, from the flight plan or from the aircraft, that is less than a predetermined maximum distance and of which at least one period of validity verifies a predefined display time criterion, the said extraction means transmitting the said extracted datum or data to the said navigation display capable of displaying the representations of the said extracted data.

The invention discloses a method and equipment for classifying audio signals in real time, belonging to the technical field of audio encoding and decoding. The method comprises the following steps: preprocessing the inputted audio signals; extracting multi-level audio features in a time domain and an MDCT domain; classifying a current frame by adopting the signal audio feature if the current frame is in a classification convergence time frame I; classifying the current frame by adopting the coarse-fine classification rule if the current frame is behind the classification convergence time frame I, and judging classification features by adopting the multi-level features; and after classifying the current frame by adopting the coarse-fine classification rule, upgrading the signal classification type of the current frame according to the historical state of the classification type of a signal frame before the current frame. Due to the adoption of the method and the equipment, the signals can be classified in real time simply and accurately.

The invention relates to a method and system of electrocardiosignal classification. The method comprises the steps of obtaining electrocardiographic test data, processing the electrocardiographic test data to obtain a waveform graph with the predetermined size, and combined with the oscillogram with the predetermined size and an established heart disease classification type model, determining a classification result corresponding to the electrocardiographic test data, wherein the heart disease classification type model is a model capable of reflecting original waveform characteristics and the classification result. By utilizing the method of electrocardiosignal classification, the accuracy rate of the classification result can be increased, and the purpose that the heart rate variability of more types is classified can be achieved at the same time.

The invention provides a fault / potential hazard knowledge base system and an establishment method. The method comprises the steps of classifying common faults and potential safety hazards of equipment, and establishing codes according to a quantity of classification types; and automatically judging fault reasons, fault positions, repair measures and the like of the equipment according to the fault / potential hazard knowledge base system, and when the fault reasons, the fault positions, the repair measures and the like of the equipment are inconsistent with preset fault mode or potential hazard phenomenon information, adding fault modes or potential hazard phenomena in a manual input mode and inputting the added fault modes or potential hazard phenomena to an equipment fault and potential safety hazard knowledge base. The centralized storage of fault information / potential hazard check standard data is realized; a structured collection of information resources is formed; a fault / potential hazard knowledge base based on universal equipment / industry classification is formed; the structured management of related knowledge experience is facilitated; the value of equipment management and maintenance knowledge is extended; the accuracy and working efficiency of identification and maintenance are improved; and the waste of social information resources is avoided.

The invention relates to the field of data processing and machine learning classification, in particular to a method for optimizing word classification in machine learning text. The method comprises the steps that on the basis of text classification, self-defined and semantically related features are filtered out through a feature selection regulator based on a regular expression, a user customizes classification types in training data after feature selection, and classification based training is conducted by means of the features and the types according to a naive Bayesian model; after training is completed, in the application stage, if statements conforming to the feature selection regulator exist in text needing word classification, classification is completed by combining a trained model. According to the method, the capacity of the model for processing work classification is not limited in word data in a training sample; the method can be applied to optimization and application of machine learning text work classification and derivation functions thereof.

The invention provides an intelligent garbage sorting device, a terminal and an intelligent garbage sorting system. The intelligent garbage sorting device comprises a garbage throwing opening, a detection component and a conveying passage, wherein the garbage throwing opening is used for receiving current thrown garbage; the detection component is communicated with the garbage throwing opening, and is used for obtaining the attribute information of the current thrown garbage and determining the classification types of the current thrown garbage according to the attribute information; and the conveying passage is connected with the detection component, and is used for conveying the current thrown garbage in corresponding accommodating devices according to the classification types of the current thrown garbage. Through the technical scheme, the classification types of the garbage can be determined intelligently, so as to realize the automatic sorting of the garbage, meet the requirement of the garbage sorting advocated by the modern society, increase the recovery rate of garbage resources, and improve the user experience.

The present invention is applied to the electric power system field, in particular relates to a gray prediction and support vector machine-based classification type electric vehicle demand temporal-spatial distribution dynamic prediction method. The method comprises the steps of firstly utilizing a high-accuracy improved gray model to predict the numbers of the vehicles of different types; then based on the proportions of the electric vehicles of different types and the nonlinear characteristic of the influence factors, utilizing a support vector machine regression method and the prediction samples to obtain a classification type electric vehicle substitution ratio prediction result, and utilizing an iterative method to correct the prediction result continuously; and finally matching the two prediction results according to the vehicle types, establishing a classification type electric vehicle demand growth prediction model, and combining the user travel rule research to realize the accurate dynamic temporal-spatial prediction of the classification type electric vehicle demands. Therefore, the method of the present invention has the following advantages of fully considering the characteristic that the historical data is insufficient and the influence of different factors on the development of the electric vehicles and combing the user travel rule research to realize the more accurate dynamic prediction.

The invention relates to the technical field of reservoir research, and discloses a reservoir classification method and system. The method includes the steps that firstly, a mercury injection experiment is conducted on a rock core to acquire mercury injection experimental data of the rock core, and then pore structures of the rock core are classified according to the mercury injection experimental data of the rock core to acquire classification types of the pore structures of the rock core; standard mercury injection coefficients are calculated through a formula; the classification types of the pore structures of the rock core are combined with the standard mercury injection coefficients to acquire classification types of comprehensive parameters of reservoirs; scales of the rock core are utilized to conduct logging, and actual mercury injection coefficients are extracted from a logging curve; the classification types of the comprehensive parameters of the reservoirs are combined with the actual mercury injection coefficients to classify the reservoirs, and then classification types of the reservoirs are acquired. By the adoption of the method, continuous classification of the reservoirs within a conventional logging depth reservoir range is achieved, classification efficiency of the reservoirs is improved, and reliable reservoir classification parameters and bases are provided for evaluation of large-area reservoir exploration and petroleum reservoir exploitation of China at present.

The disclosed technology includes a distributed ledger system that can include one or more transaction blocks that are associated with one or more transaction records of an account corresponding to the primary account number. The system may allocate a first type block in a first distributed ledger, and the first type block may be utilized to identify a classification type associated with the first primary account number. The system may receive, at the first node, a first transaction indication associated with the first primary account number. Responsive to receiving the first transaction indication, the system may bind a first transaction block to the first distributed ledger. Transaction blocks may be configured for selective control and / or selective replicated to other nodes of the system. In response to an end-of-cycle indication, the first transaction block may be locked to prevent further modification.

In an example embodiment, a method for selecting a communication path is provided. The method may comprise receiving data encapsulated in a transport protocol. In addition, a classification type and exit path information associated with the classification type may be received. The data is associated with the classification type and then is encapsulated in Stream Control Transmission Protocol (SCTP) based on the exit path information. This exit path information is associated with the classification type that is associated with the data.

A laser-point cloud and image fusion data classification method based on multi-characteristic comprises the following steps: 1, data preprocessing: preprocessing aviation image and unmanned plane laser-point cloud data; 2, sample extraction: fully utilizing geometry characteristics of the point cloud data and spectrum characteristics of the aviation images so as to carry out sample extraction of various types; 3, fusion data classification based on multi-characteristic: using a vector description model to classify the sample data; 4, precision evaluation: evaluating the precision of the classified data. The method is complete in surface object extraction and high in classification precision. The method starts from the angle of the fusion image spectrum information, carries out data fusionaccording to application purposes and surface object classification demands, sets corresponding classification rules for classification of main surface objects, and builds a corresponding relation between classification types and classification characteristics, thus extracting complete surface object areas, and reducing misclassification phenomenon.

The invention discloses a method and a device for text classification, and a method and a device for characteristic processing of text classification. The method for characteristic processing of text classification includes obtaining a characteristic set of learning materials used for text classification; calculating a sum of information gain values of each characteristic word in all classification types; and extracting characteristic words in a preset number in the characteristic set as learning characteristics used for text classification to enable the learning characteristics used for text classification to be part of the characteristic words in residual characteristic words except for stop words in the characteristic set, wherein the sum of the information gain values corresponding to extracted characteristic words is larger than the sum of the information gain values corresponding to non-extracted characteristic words. By applying the method and the device for text classification to characteristic extraction of text classification, noise characteristics can be effectively avoided being brought into a machine learning process, so that accuracy of text classification is improved, the scale of a characteristic library is greatly reduced, and memory usage is reduced.

The invention discloses a content-based image recognition searching method, a mobile terminal, a server and a system, wherein the mobile terminal is used for shooting an image of an unknown object, sending the shot image of the unknown object to the server at a cloud side after a plurality of target regions with rich information contents or the classification types of the unknown object are selected, finally receiving inquiry result information pushed by the server and performing display; the server is used for receiving the unknown object image sent by the mobile terminal, performing feature classification to extract image feature values, then, performing searching in an image index relationship table of the unknown object according to the image feature values, and finally pushing the searched unknown object image and the unknown object information to the mobile terminal. The content-based image recognition searching method disclosed by the invention has the advantages that the object searching convenience and accuracy under the condition of only knowing the appearance of the unknown object are improved; high application prospects are realized.

The invention discloses a method for identifying winter wheat. On the basis of remote-sensing vegetation index time sequence data and phonological calendar data during a growth period of winter wheat, classification type setting, training sample selection, sample time sequence curve extraction, winter wheat identification feature selection and parametrization, optimal feature threshold determination and winter wheat identification model construction are carried out successively to extract spatial distrubtion information of the winter wheat in a range. The method is characterized in that spatial distribution information extraction of the winter wheat can be carried out based on the remote-sensing vegetation index time sequence data without any de-noising processing, so that the anti-time sequence data noise capability is high. The method has high stability and universality can be applied to remote-sensing extraction of distribution information of the winter wheat in a large range having a certain phonological difference.

An information processing apparatus, such as might be associated with an in-home audio or video control device or a home appliance operated by a user, is provided with input allowing the user to set accuracy of a log of device operation history and / or to set a ratio of disclosure of the log information. A log accumulation control device efficiently collects the log information and offers users diversified levels of accuracy for the log information. In a further configuration, user-designated information acquired form user operation of the device identifies frequency of viewing and listening to each program, and reproduction of contents is controlled as a function of the user-designated information, e.g. to select among common contents and contents of different classification types.

The invention relates to a classification type efficient ultrafine grinder comprising a rack (1), wherein a feeding chamber (18), a grinding chamber (14) and a discharging chamber (15) are arranged on the rack (1); a grading chamber (23) is arranged at the top of the rack (1) and is located above the discharging chamber (15); the discharging chamber (15) is connected with the grading chamber (23) through a conveying pipeline I (33); a coarse material discharging opening (34) is formed in the bottom of the grading chamber (23); a coarse material air seal machine (22) is arranged below the coarse material discharging opening (34) in a connected manner; and the lower end of the coarse material air seal machine (22) is connected with a feeding opening (21) of the grinding chamber (14). According to the classification type efficient ultrafine grinder provided by the invention, a material enters the grinding chamber for three grinding processes; the grinded material is fed into the grading chamber; the material with a required fineness enters a cyclone discharging process; and a coarse material falls back to the grinding chamber for secondary grinding, so that the heat of a machine cavity is effectively reduced, and the production efficiency is improved.

The invention discloses an automatic filing system and method for court electronic files. A computer automated processing technique is used, the image feature values of the electronic file titles arelearned, trained and recognized through the artificial intelligence technology and matched with the preset feature library, the functions of automatic recognition and automatic filing of the electronic files are achieved, new types are continuously learned through a learning training model to obtain new classification feature values, and therefore the feature library is continuously updated to enrich the classification types. According to the invention, the computer and artificial intelligence are combined, and manual intervention is taken as assistance to replace an original manual processingmode to archive and classify the electronic files of the court, so that the working efficiency and the archiving accuracy are effectively improved; and through the learning training model, the computer continuously learns to adapt to more types of files, so that the filing classification is further perfected, and the recognition range of the system and method is expanded.

The invention discloses a facial skin quality evaluation method, which comprises the steps of collecting a facial image; graying the facial image through an open source library, and determining coordinates of the face; constructing a mask image of a facial skin region; segmenting the facial image; performing preprocessing and feature extraction; training a classification model of a support vectormachine (SVM) with the classification type being C_SVC, the kernel type being polynomial kernel POLY and the kernel parameter degree being 1.0 so as to obtain a trained SVM classifier; marking the recognized facial acne position with a black circle, and marking the recognized wrinkle position with a black box; and evaluating the skin quality according to the number of the recognized acnes and thefacial proportion occupied by the wrinkles. The facial skin quality evaluation method can eliminate the subjectivity of manual evaluation by judging after learning. Meanwhile, the facial skin qualityevaluation method unifies the brightness, performs Gamma correction and graying, extracts original LBP features and carries out histogram equalization, thereby reducing the calculation amount, and improving the real-time performance and the self-adaptation.

The embodiment of the invention provides a data prediction method and device, electronic equipment and a storage medium. The method comprises the following steps: obtaining feature information of to-be-predicted data and scene information of the to-be-predicted data, inputting the acquired information into a prediction model; obtaining prediction results, wherein the prediction model is obtained by training the following steps: acquiring scene information of the sample data, feature information of the sample data and classification types of the sample data; and inputting the scene informationof the sample data and the feature information of the sample data into a to-be-trained FFM model to obtain a first numerical value corresponding to the sample data, determining parameters of the to-be-trained FFM model according to the first numerical value and the classification category of the sample data, and determining a prediction model according to the to-be-trained FFM model after the parameters are determined. According to the embodiment of the invention, the data of different scenes can be predicted by training one model, and the models do not need to be trained for multiple scenes respectively, so that the model training process is simplified, the model training efficiency is improved, and the data prediction efficiency is higher.